1Z0-1072-25 Question 8
Single answerYour analytics team needs to run high-throughput, batch data processing jobs on Oracle Cloud Infrastructure. These jobs require large amounts of temporary scratch space to store intermediate data and must handle high IOPS to avoid bottlenecks. You plan to spin up compute instances only during processing windows and shut them down afterward to minimize costs. Which compute shape would best meet these requirements?
- A
VM.Standard3.Flex
- B
BM.DenseIO2.52
- C
HPC2.36
- D
BM.GPU3.8
Show answer and explanation
Correct answer: B
Explanation
For workloads requiring large, high-performance ephemeral storage, DenseIO shapes on OCI offer local NVMe SSDs with excellent IOPS and throughput. These are ideal for temporary data storage and batch analytics tasks that need quick disk operations. According to Oracle documentation on Compute shapes, BM.DenseIO shapes provide local NVMe storage, making them the best choice for high-throughput data processing with large intermediate datasets.
- A. Incorrect.
Option 1 (VM.Standard3.Flex): While flexible shapes allow you to choose the number of OCPUs and memory, they do not provide large amounts of local NVMe storage for high IOPS. You would need to rely on block volumes, which may not deliver the same performance for temporary scratch space.
- B. Correct.
Option 2 (BM.DenseIO2.52): This bare metal shape includes high-performance NVMe SSDs with substantial local storage, delivering very high IOPS. It is ideal for compute jobs requiring large amounts of ephemeral data storage and quick read/write operations, making it the best fit for the scenario.
- C. Incorrect.
Option 3 (HPC2.36): HPC shapes are optimized for high-performance computing tasks, especially those requiring low-latency cluster networking. Although they provide strong compute capabilities, they do not focus on large local NVMe storage for big data workloads. Therefore, they are less suitable for this scenario where local storage performance is crucial.
- D. Incorrect.
Option 4 (BM.GPU3.8): This shape is targeted at GPU-intensive workloads such as machine learning or rendering. While it includes powerful GPUs, it is not the most cost-effective or optimized choice for high-volume data processing using local storage.